In human-intensive business processes such as IT incident management, failures and/or issues are commonly reported by customers as service requests. Service groups can analyze such requests and write observations in the form of communication logs. Such logs are commonly manually summarized by service practitioners as problem tickets. Communication logs can capture details of service groups that logged observations, such as, for example, the name of an owner group, a time-stamp, and the order of owner groups in which observations are logged. Ticket analysis can help create problem diagnostics, as well as predict outages and prevent one or more issues. However, practitioners can often write incomplete, incorrect or abbreviated observations. Such flaws in ticket descriptions can render ticket analysis challenging and inefficient.